Entrapment and dissolution of DNAPLs in heterogeneous ......homogeneous media with cylindrical...
Transcript of Entrapment and dissolution of DNAPLs in heterogeneous ......homogeneous media with cylindrical...
www.elsevier.com/locate/jconhyd
Journal of Contaminant Hydrology 67 (2003) 133–157
Entrapment and dissolution of DNAPLs in
heterogeneous porous media
Scott A. Bradforda,*, Klaus M. Rathfelderb,John Langb, Linda M. Abriolab
aGeorge E. Brown, Jr., U.S. Salinity Laboratory, USDA, ARS, 450 W. Big Springs Road,
Riverside, CA 92507-4617, USAbDepartment of Civil and Environmental Engineering, University of Michigan,
181 EWRE, 1351 Beal Avenue, Ann Arbor, MI 48109-2125, USA
Received 9 October 2001; accepted 28 March 2003
Abstract
Two-dimensional multiphase flow and transport simulators were refined and used to numerically
investigate the entrapment and dissolution behavior of tetrachloroethylene (PCE) in heterogeneous
porous media containing spatial variations in wettability. Measured hydraulic properties, residual
saturations, and dissolution parameters were employed in these simulations. Entrapment was
quantified using experimentally verified hydraulic property and residual saturation models that
account for hysteresis and wettability variations. The nonequilibrium dissolution of PCE was
modeled using independent estimates of the film mass transfer coefficient and interfacial area for
entrapped and continuous (PCE pools or films) saturations. Flow simulations demonstrate that the
spatial distribution of PCE is highly dependent on subsurface wettability characteristics that create
differences in PCE retention mechanisms and the presence of subsurface capillary barriers. For a
given soil texture, the maximum and minimum PCE infiltration depth was obtained when the sand
had intermediate (an organic-wet mass fraction of 25%) and strong (water- or organic-wet)
wettability conditions, respectively. In heterogeneous systems, subsurface wettability variations were
also found to enhance or diminish the performance of soil texture-induced capillary barriers. The
dissolution behavior of PCE was found to depend on the soil wettability and the spatial PCE
distribution. Shorter dissolution times tended to occur when PCE was distributed over large regions
due to an increased access of flowing water to the PCE. In heterogeneous systems, capillary barriers
that produced high PCE saturations tended to exhibit longer dissolution times.
Published by Elsevier B.V.
Keywords: Nonaqueous phase liquid (NAPL); Dissolution; Entrapment;Wettability; Heterogeneity; Interfacial area
0169-7722/$ - see front matter. Published by Elsevier B.V.
doi:10.1016/S0169-7722(03)00071-8
* Corresponding author. Tel.: +1-909-369-4857; fax: +1-909-342-4963.
E-mail address: [email protected] (S.A. Bradford).
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157134
1. Introduction
The improper storage and disposal of nonaqueous phase liquids (NAPLs) has resulted
in widespread contamination of the subsurface environment (National Research Council,
1994). NAPLs migrate downward through the heterogeneous subsurface under the
influence of gravity and capillary forces. Upon reaching the saturated zone, an NAPL
may move laterally within the capillary fringe zone at the water table or, for denser than
water NAPLs (DNAPLs), the saturated zone may not act as an impediment to vertical
migration and the DNAPL may continue to move downward. As the NAPL migrates,
capillary forces act to retain a residual portion of the mass as ganglia in the center of the
pore space (Chatzis et al., 1983) or as films coating NAPL-wet solids. In soil column
experiments, entrapped residual NAPL saturations typically range from 5% to 35% of the
pore space (Wilson et al., 1990; Powers et al., 1992, 1994a,b; Bradford et al., 1999). When
an infiltrating NAPL encounters an interface of capillary contrast, the NAPL tends to
accumulate in pools and spread laterally along such interfaces (Wilson et al., 1990;
Bradford et al., 1998). If such an interface is tilted, the infiltrating fluid will be diverted
and flow down the interface as capillary diversion. Capillary barriers occur naturally in
heterogeneous systems due to differences in soil texture (e.g., Kueper et al., 1989, 1993;
Wilson et al., 1990; Kueper and Frind, 1991; Poulsen and Kueper, 1992; Essaid et al.,
1993; Dekker and Abriola, 2000) and/or porous medium wettability characteristics
(Bradford et al., 1998). Downward migration of NAPL will continue in the subsurface
until the NAPL is completely retained as entrapped residual or pooled NAPL. This final
distribution of NAPL is referred to as the NAPL source zone.
The slow partitioning of NAPL in the source zone to flowing water serves as a
persistent source of groundwater contamination. Regulated exposure levels for many
NAPLs are several orders of magnitude below their solubility limits. Subsurface NAPL
contamination is therefore a problem of serious concern at hazardous waste sites.
Experimental studies have been undertaken to investigate the dissolution behavior of
uniformly distributed residual NAPL in soil columns (e.g., Hunt et al., 1988; Miller et al.,
1990; Parker et al., 1991; Guarnaccia et al., 1992; Powers et al., 1991, 1992, 1994a,b;
Imhoff et al., 1994, 1997; Rixey, 1996). Most of these studies have been conducted using
water-wet glass beads or silica sands. Results from these studies indicate that dissolution
behavior depends on system hydrodynamics, NAPL saturation and composition, and soil
grain-size distribution characteristics. However, these studies have not produced clear
consensus on the effect of soil grain-size distribution parameters and NAPL saturation on
dissolution behavior (cf., Miller et al., 1998; Bradford et al., 2000).
Laboratory-scale dissolution experiments have also been conducted in larger-scale
systems (Anderson et al., 1992; Geller and Hunt, 1993; Powers et al., 1998; Saba and
Illangasekare, 2000). Both Geller and Hunt (1993) and Anderson et al. (1992) employed
homogeneous media with cylindrical regions of residual NAPL blobs embedded in the
center. Geller and Hunt (1993) attributed reductions in the dissolution mass transfer rate to
the reduced effective permeability within the region of residual NAPL. In contrast,
Anderson et al. (1992) concluded that reductions in dissolution mass transfer rate in their
experiments were attributed to dispersion and dilution of contaminated water. Saba and
Illangasekare (2000) reported on two-dimensional dissolution experiments in which the
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 135
residual NAPL was entrapped in rectangular shaped source zones of various lengths. They
found that the experimental lumped mass transfer coefficient increased with increasing
size of the contaminant source zone. Powers et al. (1998) conducted dissolution experi-
ments in heterogeneous systems consisting of NAPL entrapped at high saturation in
coarse-textured lenses. They found that initial dissolution rates could be adequately
predicted by accounting for system hydrodynamics (variations in intrinsic and effective
permeabilities). At low NAPL saturations, however, rate-limited NAPL dissolution was
found to become important.
Numerical models have been used to predict the entrapment and dissolution behavior of
NAPLs in heterogeneous porous media (Mayer and Miller, 1996; Berglund, 1997; Dekker
and Abriola, 2000). These studies indicate that increasing the variance of the hydraulic
conductivity distribution (physical heterogeneity) tends to increase the dissolution time.
Correlations for the lumped mass transfer coefficient incorporated into these models were
developed from small-scale soil column experiments with uniform residual NAPL
saturations. Application of these relationships to heterogeneous systems often requires
extrapolation to much higher NAPL saturations and porous media characteristics beyond
the scope of experimental conditions considered in correlation development. Not surpris-
ingly, Mayer and Miller (1996) reported that differences in the predicted dissolution
behavior for various mass transfer coefficient correlations increased with increasing degree
of physical heterogeneity.
In natural subsurface systems, highly irregular NAPL distributions have been observed
(Poulsen and Kueper, 1992; Essaid et al., 1993). Variations in soil texture and wettability
may account for some of these irregularities. In comparison to information on soil texture,
there is a paucity of data on subsurface wetting properties and no systematic studies of
NAPL-water contact angle variations have been reported. Heterogeneity of subsurface
wettability characteristics may occur at hazardous waste sites as a result of variations in
aqueous chemistry (Demond et al., 1994), mineralogy (Anderson, 1986), organic matter
distributions (Dekker and Ritsema, 1994), and contaminant aging (Powers and Tamblin,
1995). Bradford et al. (1999) presented soil column entrapment and dissolution data for
tetrachloroethylene (PCE) in soils representative of a range of grain-size and fractional
wettability characteristics. Organic entrapment and dissolution rates were found to depend
strongly on wettability and grain-size distribution characteristics. These observations
suggest that the superposition of soil texture and wettability properties in heterogeneous
environments will likely determine the entrapment and dissolution behavior of NAPLs. To
date, relatively little research has explored the influence of subsurface wettability
variations on NAPL distribution and remediation.
The objective of this work is to explore the entrapment and dissolution behavior of
DNAPLs in porous media containing spatial variations in soil texture and wettability.
Conceptual models for NAPL entrapment and dissolution that were developed from
experimental data (Bradford and Abriola, 2001) were implemented into existing two-
dimensional NAPL flow (Abriola et al., 1992) and transport (Abriola et al., 1997;
Rathfelder et al., 2000) simulators. The entrapment model presented by Bradford et al.
(1998) was extended to account for the nonlinear influence of wettability on NAPL
entrapment. The dissolution model presented by Bradford and Abriola (2001), which
utilizes independent estimates of the film mass transfer coefficient and interfacial area, was
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157136
modified to account for both entrapped and continuous (immobile films and NAPL pools)
NAPL saturations. Independently determined hydraulic properties, residual saturations,
and dissolution parameters for soils encompassing a range of grain-size and wettability
characteristics are used herein, in conjunction with the refined numerical simulators.
Emphasis is placed upon NAPL recovery; sorption/desorption processes are not consid-
ered in the simulations.
2. Models of NAPL entrapment and dissolution
2.1. Entrapment
MVALOR, a two-dimensional three-phase immiscible flow simulator, is employed
herein to investigate the entrapment behavior of PCE in saturated systems. A description
of the numerical methods, model verification and testing, and example model applications
for this simulator are available in the literature (Abriola et al., 1992; Abriola and
Rathfelder, 1993; Demond et al., 1996; Rathfelder and Abriola, 1998; Bradford et al.,
1998; Dekker and Abriola, 2000). In this work, the model will be used to solve the
coupled mass conservation equations for the NAPL and aqueous phases (e.g., Abriola,
1989):
B
BtðeqiSiÞ ¼ r k
qikri
li
ðrPi � qigrzÞ� �
þ qiRi ð1Þ
where S is the saturation, P (M L� 1 T� 2) is the pressure, e is the porosity of the medium, l(M L� 1 T� 1) is the viscosity, k (L2) is the intrinsic permeability, q (M L� 3) is the density,
g (L T� 2) is the acceleration due to gravity, R (T� 1) is the source/sink term, z (L) is the
positive downward vertical direction, j is the gradient operator (B/Bx, B/Bz; x is the
horizontal direction), and the subscript i denotes the distinct fluid phases (i.e., i = o,w for
organic and water, respectively). These mass conservation equations require that
So + Sw = 1.
To account for the influence of wettability on NAPL migration, hysteretic fractional
wettability hydraulic property models (Bradford and Leij, 1996; Bradford et al., 1997) that
were developed from careful comparison to experimental data were incorporated into
MVALOR (Bradford et al., 1998). Measured capillary pressure (Pow)–saturation data
were described with the modified van Genuchten model (Bradford and Leij, 1995a) (cf.
Table 1). Relative permeability relations were estimated from capillary pressure data
according to the pore-size distribution model presented by Bradford et al. (1997). This
model accounts of the influence of wettability on the relative permeability relations. The
model predicts that for a given saturation and saturation history, an increasing organic-wet
fraction tends to decrease the NAPL relative permeability and increase the water relative
permeability.
In fractional wettability porous media, the residual water (Srw) and NAPL (Sro)
saturations consist of the sum of immobile and entrapped saturations. Values of Srw and
Sro that are utilized herein were obtained from measured capillary pressure–saturation
Table 1
Hydraulic properties of experimental systems
Fo e n ad (cm� 1) ai (cm
� 1) kd (cm) ki (cm) Srw Sro k (m2)
F20-F30
0.00 0.315 5.875 0.124 0.264 0.000 0.000 0.159 0.100 4.08e� 10
1.00 0.328 5.875 0.124 0.264 9.570 10.00 0.100 0.065 4.08e� 10
F35-F50
0.00 0.313 5.359 0.055 0.129 0.000 0.000 0.040 0.200 6.37e� 11
0.25 0.313 5.359 0.055 0.129 6.740 6.090 0.014 0.109 6.37e� 11
0.50 0.313 5.359 0.055 0.129 14.30 12.50 0.045 0.100 6.37e� 11
0.75 0.315 5.359 0.055 0.129 18.87 17.23 0.084 0.066 6.37e� 11
1.00 0.321 5.395 0.055 0.129 20.44 22.58 0.050 0.050 6.37e� 11
F70-F110
0.00 0.331 9.264 0.022 0.044 0.000 0.000 0.245 0.245 4.68e� 12
0.25 0.325 9.264 0.022 0.044 20.06 17.64 0.080 0.080 4.68e� 12
0.50 0.330 9.264 0.022 0.044 37.32 40.30 0.080 0.110 4.68e� 12
1.00 0.342 9.264 0.022 0.044 59.67 66.51 0.180 0.070 4.68e� 12
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 137
data. An immobile wetting fluid residual occurs due to the presence of thin films (water or
NAPL) coating solid surfaces, whereas entrapped nonwetting fluid residual occurs in
larger portions of the pore space as a result of capillary instabilities as wetting fluid
invades (Chatzis et al., 1983). Bradford et al. (1998) postulated that entrapped (Sot) and
immobile (Soi) residual NAPL saturations can be linearly related to the NAPL-wet sand
mass fraction (Fo) as (1�Fo)Sro and FoSro, respectively. Recent analysis of measured
residual NAPL saturation values and dissolution behavior in fractional wettability media,
however, suggests that the magnitude of Sot and Soi vary in a more complex and nonlinear
manner with Fo and soil grain-size distribution (Bradford et al., 1999). It is assumed herein
that entrapped and immobile residual NAPL saturations are given as xo Sro and
(1�xo)Sro, respectively. Here, xo is used to account for the nonlinear influence of
fractional wettability on entrapped and immobile residual saturations. Based upon the
fitted values of xo to fractional wettability dissolution data (developed from 20 soil
column experiments using various-sized Ottawa sands, including those listed in Table 1),
the following empirical correlations were developed by Bradford and Abriola (2001):
xo ¼ ð1� FoÞ11:44 d50 < 0:071
xo ¼ ð1� FoÞ42:79 d50z0:071 ð2Þ
Values of the entrapped (Swt) and immobile (Swi) water saturation were determined
analogously to Sot and Soi by reversing the roles of NAPL and water. When Sot =xo Sro, it
is assumed that all the accessible pore space has contacted NAPL. To account for
hysteretic effects on Sot, the procedures of Land (1968) and Parker and Lenhard (1987)
are employed. Hysteretic effects on Soi are modeled by multiplying (1�xo)Sro by Somax/
(1� Swi); here Somax is the historic maximum NAPL saturation. This approach assumes
that Soi is proportional to the drainage pore space that is accessible to the NAPL.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157138
A qualitative experimental validation of MVALOR was recently provided by O’Carroll
et al. (2001). These authors presented a comparison between measured and MVALOR-
simulated PCE infiltration and redistribution in a two-dimensional heterogeneous system
containing spatial variations in soil texture (Ottawa sands given in Table 1) and wettability.
The MVALOR model results compared quite well with the experimental data when using
independently determined hydraulic properties.
2.2. Dissolution
The MISER model is employed herein to investigate the dissolution behavior of PCE
in saturated systems. A description of the numerical methods, model verification and
testing, and example model applications for this simulator are available in the literature
(Abriola et al., 1997; Rathfelder et al., 2000, 2001). MISER currently solves a coupled
system of two-phase flow (air–water) and multicomponent transport/reaction equations,
subject to rate-limited mass transfer. The model also incorporates an immobile NAPL
phase. In MISER, a transport equation for a given component in the aqueous phase is
modeled as
B
BtðeSwCÞ ¼ rðeSwDh � rCÞ � rðqwCÞ þ
Xg
Egw þ B ð3Þ
where qw (LT� 1) is the Darcy velocity vector,Dh (L2 T� 1) is the component hydrodynamic
dispersion coefficient tensor, C (M L� 3) is the component concentration, B (M T� 1 L� 3) is
the net rate of biological transformation of the component, and Egw (M T� 1 L� 3) is the
interphase mass transfer rate of the component across the aqueous and g phase interface. A
similar equation is written for the component mass balance equation for the NAPL phase. In
this case, however, the advective and dispersive solute fluxes are set equal to zero.
Nonequilibrium dissolution is accounted for in MISER using an quasi-steady state
approximation of Fick’s first law of diffusion. The dissolution model presented below is an
adaptation of the model presented by Bradford and Abriola (2001) that was developed to
describe the soil column dissolution behavior of residual NAPL entrapped in fractional
wettability porous media. This model is modified below to explore the dissolution
behavior of NAPLs in heterogeneous systems. In such systems, the final NAPL saturation
distribution from MVALOR serves as the initial condition for the MISER simulations. The
NAPL saturation (So) at a particular location is assumed to be the sum of contributions
from NAPL entrapped as ganglia in water-wet pores (Sot) and as continuous NAPL phase
(Soc) occurring as immobile films coating NAPL-wet solid surfaces or free product NAPL
pools. The determination of Sot was discussed above in Section 2.1. The value of Soc is
determined as So� Sot.
The following expression is employed herein to describe the mass exchanged per unit
time per unit volume between the NAPL and aqueous phases (Eow) (Bradford and Abriola,
2001):
Eow ¼ �BðeSoqoÞBt
¼ kowðaAtow þ bAc
owÞðCe � CÞ ð4Þ
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 139
Here kow (L T� 1) is known as the film mass transfer coefficient, Aowt (L� 1) is the
NAPL-water interfacial area per unit volume of porous media attributed to entrapped
ganglia, Aowc (L� 1) is the NAPL-water interfacial area per unit volume of porous media
attributed to the continuous NAPL phase, Ce (M L� 3) is the equilibrium solubility
concentration that approximates the concentration at the NAPL-water interface, and a and
b are parameters to account for the fact that only a fraction of the interfacial area is
exposed to mobile water.
Independent estimates for kow and the NAPL-water interfacial area are utilized herein.
The value of kow is estimated using a slightly modified form of the dimensionless
correlation developed by Powers et al. (1994b). Powers et al. (1994b) measured the
dissolution behavior of solid naphthalene spheres, with a known size and interfacial area
that had been embedded in sandy porous media. Measured values of kow from this study
were assumed to be independent of NAPL configuration because the interfacial area was
well characterized. These authors established a dimensionless power function correlation
that relates kow measurements to experimental porous medium (median grain size) and
hydrodynamic (aqueous phase pore water velocity) characteristics. Bradford and Abriola
(2001) modified this expression to account for the reported dependence of kow on NAPL
diffusivity and viscosity (Imhoff et al., 1997).
The influence of NAPL configuration on dissolution behavior is explicitly modeled
herein by accounting for temporal changes in the NAPL-water interfacial area. The
determination of interfacial area attributed to the entrapped ganglia (Aowt ) was recently
discussed in detail by Bradford and Abriola (2001). At a particular location, entrapped
NAPL is initially distributed among a number of ganglia classes. Each ganglia class
consists of spherical singlets associated with a different saturation-dependent radius and
NAPL saturation (Powers et al., 1991, 1994a,b). The initial values of saturation-dependent
singlet class radii are estimated from NAPL-water capillary pressure (Pow)–water
saturation (Sw) data using Laplace equation for capillarity and assuming that the ganglia
are entrapped in the largest portions of the pore space (i.e., in the saturation range
1� Sot < Sw < 1) (Bradford and Leij, 1997). The initial interfacial area per unit volume for
each pore class is then estimated as the product of the surface area of the sphere and the
number of spheres per unit volume. Bradford and Abriola (2001) reported favorable
agreement between Aowt estimates and measurements (entrapped styrene ganglia that had
been polymerized in situ, removed from the soil, and then sieved into various ganglia size
classes). Both Powers et al. (1994b) and Bradford and Abriola (2001) have used a
distribution of NAPL spheres to successfully model the long-term dissolution behavior of
entrapped NAPL ganglia.
The value of interfacial area attributed to the continuous NAPL phase, Aowc , at a given
water saturation was estimated from Pow–Sw data. Thermodynamic considerations suggest
that the area under a capillary pressure curve is related to the NAPL-water interfacial area
(Leverett, 1941; Morrow, 1970; Bradford and Leij, 1997). In this work, an estimate of Aowc
for a particular porous medium is obtained from main imbibition Pow–Sw data as
Acow ¼ � e
row
Z Sw
vPowðSÞdS ð5Þ
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157140
Here the limit of integration v is assumed to be the value of Sw, where Pow = 0, row
(MT� 2) is the interfacial tension, and S is a dummy saturation variable of integration. In
water-wet systems, the value of Pow is exclusively positive and v equals 1� Sot. In
contrast, in fractional wettability systems, the value of Pow can be positive or negative
depending on the water saturation and v tends to decrease with increasing organic-wet
solid fraction. Bradford and Abriola (2001) estimated the NAPL-water interfacial area of
residual NAPL films coating NAPL-wet solid surfaces in a slightly different manner as the
product of the area of drained pore space and the organic-wet mass fraction; Eq. (5) is
believed to provide a better description of temporal changes in Aowc .
The above relations provide initial estimates of the interfacial area for a given NAPL
saturation. Temporal changes in interfacial area as a result of dissolution are modeled
according to the approach presented by Bradford and Abriola (2001). As dissolution
proceeds, changes in NAPL saturation during a given time interval (DSo) are evaluated as
the sum of changes in the entrapped (DSot) and continuous (DSoc) NAPL saturations. Eq.
(4) indicates that the values of DSot and DSoc will be related to the magnitudes of
(aAowt DSo)/(aAow
t + bAowc ) and (bAow
c DSo)/(aAowt + bAow
c ), respectively. Temporal changes in
Aowc for a particular location are calculated with Eq. (5). Hence, Aow
c is a function of Swuntil Soc is zero, at which time Aow
c is set equal to zero. The value of Aowt is also modeled as
an explicit function of saturation. In this case, the ganglia radii slowly decrease in size as
Sot decreases. The saturation dependence of a particular ganglia class radius is related to
the initial ganglia class radius and NAPL saturation (Bradford and Abriola, 2001). The
interfacial area is then calculated as the product of the surface area of the sphere and the
number of spheres per unit volume. Note that only initial ganglia class radii are estimated
using capillary pressure data.
Values of a and b are predicted herein from correlations established by Bradford and
Abriola (2001) between fitted values and experimental parameters (developed from 20 soil
column experiments using various-sized Ottawa sands, including those listed in Table 1).
The value of a was found to be inversely related to the normalized mean grain size,
suggesting that less mobile water contacts the NAPL ganglia in finer textured soils. The
value of b was found to be inversely related to the soil uniformity coefficient and Aowc .
These observations suggest that some of the continuous NAPL films are less accessible to
a flowing aqueous phase in more graded soils and that the dissolution rate depends
nonlinearly on interfacial area.
3. Numerical simulations
3.1. Entrapment
This section discusses the numerical investigation of entrapment behavior of PCE in
various subsurface systems. In this work, measured intrinsic permeability, capillary
pressure–saturation relations, and residual saturation values for various fractional wett-
ability soils are used as input to MVALOR. These experimental properties were
determined independently in soil column experiments. Fractional wettability soils were
created by combining various mass fractions of untreated and octadecyltrichlorosilane
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 141
(OTS)-treated Ottawa sands. These soils will be designated herein by their sieve sizes
(F20-F30, F35-F50, F70-F110, and F35-F50-F70-F110) and their NAPL-wet mass
fraction, Fo, or OTS percentage. The intrinsic permeability was measured using the
constant flux technique (e.g., Klute and Dirksen, 1986). The capillary pressure–saturation
relations and residual saturation were determined on similar soil columns using an
automated setup based upon the experimental design presented by Bradford and Leij
(1995b) using the pressure cell approach (e.g., Demond and Roberts, 1991). Measured
capillary pressure (Pow)–saturation data were described with the modified van Genuchten
model (Bradford and Leij, 1995a). Table 1 summarizes measured soil intrinsic perme-
abilities, capillary pressure–saturation curve parameters, and residual saturations that are
used below in numerical simulations. Here the capillary pressure parameter n is related to
the slope at the inflection point of the Pow–Sw curve (m was chosen to be equal to 1�1/n),
ad and ai are drainage and imbibition reciprocal entry pressures, and kd and ki are the
drainage and imbibition shifting parameters. The relevant physical properties of PCE that
were employed in simulations are density (1625.0 kg m� 3), viscosity (0.00089 kg ms� 1),
and interfacial tension (0.045 N m� 1).
For the simulations discussed below, PCE was introduced at the top (center node in 1D
simulations, and center 0.5 m in 2D simulations) of an initially water-saturated model
domain for a period of 1.5 days at a constant infiltration rate of 15 l day� 1 (infiltration
velocity of 1.5 and 3.0 cm day� 1 in 1D and 2D simulations, respectively). Following the
PCE release, redistribution occurred for a total simulation time of 360 days. For one-
dimensional simulations, the model domain extended 15 m in the vertical direction with a
uniform nodal spacing of 0.025 m. A no flow boundary condition for water and PCE was
enforced at the top of the model domain (PCE was added using a time-dependent source
term), while a constant pressure condition (hydrostatic as referenced to atmospheric
pressure at the domain surface) was employed at the bottom boundary. For two-dimen-
sional simulations, a 5-m (vertical) by 3.5-m (horizontal) cross-section model domain was
employed. Nodal spacing in the horizontal direction was 0.1 m. To accurately capture
capillary barrier behavior in the vicinity of various soil lens, the vertical spacing was set
equal to 0.025 m from a depth of 0.75–1.5 m, and equal to 0.05 m elsewhere in the
simulation domain. No flow boundary conditions for both water and PCE were enforced at
the top and bottom of the model domain, and hydrostatic pressure boundary conditions
were employed at the left and right boundaries (referenced to atmospheric pressure at the
domain surface). The cumulative mass balance error (Abriola et al., 1992) was less than
2� 10� 11% for the PCE entrapment simulations presented herein.
In the absence of interphase mass transfer and reactions, the final long-term NAPL
saturation distribution in homogeneous sands is controlled primarily by the values of Sro,
Sot, and Soi. Fig. 1 shows the PCE saturation distribution with depth after 360 days for a
one-dimensional F35-F50 system having organic-wet mass fractions of 0, 0.25, 0.50, 0.75,
and 1.00. Observe that higher NAPL saturation values and a lower infiltration depth occur
for the water-wet soil due to the higher Sot value in this soil. In contrast, for the other soils
(0.25, 0.50,0.75, and 1.00), the final NAPL distribution is controlled primarily by the
value of Soi. Recall that Soi increases with increasing Fo and Somax. The value of Somax in
turn is inversely related to the organic relative permeability. Since the organic relative
permeability at a given saturation decreases with increasing organic-wet mass fraction, the
Fig. 1. The PCE saturation distribution with depth after 360 days for F35-F50 porous media having organic-wet
mass fractions of 0.0, 0.25, 0.50, 0.75, and 1.00.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157142
value of Soi also increases with increasing Fo. A trend of increasing NAPL saturation and
decreasing infiltration depth is therefore observed in Fig. 1 with increasing organic-wet
mass fraction.
Analogous simulations to those shown in Fig. 1 were conducted to explore the
influence of Sro on the final NAPL saturation distribution. When the value of Sro is set
equal to 0.1, the trends observed in Fig. 1 were preserved. Changes in Sro primarily affect
the magnitude and shape of the saturation distributions. Increasing Sro tended to increase
the maximum retained NAPL saturation and therefore decreased the infiltration depth.
Fig. 2 shows the PCE saturation distribution with depth after 360 days for F70-F110
porous media having organic-wet mass fractions equal to 0.0, 0.25, 0.50, and 1.00. Here
Fig. 2. The PCE saturation distribution with depth after 360 days for F70-F110 porous media having organic-wet
mass fractions equal to 0.0, 0. 25, 0.50, and 1.00.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 143
again experimental values of Sro given in Table 1 were used in the simulations.
Comparison of Figs. 1 and 2 reveals similar trends in the maximum NAPL saturation
and infiltration depth with increasing organic-wet mass fractions. Higher NAPL satura-
tions and lower infiltration depths occur in the finer textured F70-F110 media (Fig. 2)
compared to F35-F50 sand (Fig. 1) as a result of decreases in intrinsic permeability and an
associated increase in Somax; i.e., Soi increases with decreasing intrinsic permeability.
Comparison of Figs. 1 and 2 also reveals that PCE infiltration depth is more sensitive to
organic-wet fraction as intrinsic permeability increases.
To explore the influence of spatial variations in subsurface wettability characteristics on
PCE entrapment, two-dimensional PCE infiltration and redistribution simulations were
conducted for water-wet F35-F50 porous media with various mass fractions of organic-wet
sand lenses (1 m high by 2.5 m wide) embedded in the center of the model domain. Fig.
3a–e shows the spatial distribution of PCE after 360 days when the F35-F50 sand lens has
organic-wet mass fractions equal to 0.0, 0.25, 0.50, 0.75, and 1.00, respectively. The black
box in the figure denotes the boundary of the sand lens. Note in Fig. 3a that the system is
homogeneous when the organic-wet fraction of the lens equals zero. Comparison of the
simulation results reveals that the distribution of PCE in the sand lens is highly dependent
on the wettability of the lens. In Fig. 3c–e, PCE does not displace water from the
underlying water-wet soil due to the formation of a capillary barrier at this interface; i.e.,
the PCE entry pressure of the underlying soil was not exceeded. In contrast, observe in
Fig. 3b (Fo = 0.25) that the strength of the capillary barrier is not great enough to
completely halt the migration of PCE into the underlying water-wet F35-F50 soil.
Collectively, simulation results in Fig. 3 demonstrate that contrasts in subsurface wett-
ability characteristics can create regions of high organic liquid saturation in physically
homogeneous subsurface systems.
Similar to Fig. 3, Fig. 4 presents a plot of the spatial distribution of PCE after 360 days
when various sand lenses (1 m high by 2.5 m wide) were embedded in the center of water-
wet F35-F50 sand. In Fig. 4a,b, however, the lenses are water-wet F20-F30 and F70-F110
sands, respectively. Observe that the PCE is at residual saturation directly below the
injection point. PCE pooling and spreading occurs at the top of the F70-F110 lens (Fig. 4b)
and at the bottom of the F20-F30 lens (Fig. 4a). A capillary barrier occurs at these
locations due to physical heterogeneity (pore size), and the migration of PCE from coarser
to finer sand is therefore inhibited (In Fig. 4a, from F20-F30 to F35-F50; in Fig. 4b, from
F35-F50 to F70-F110). The strength of the capillary barrier increases with increasing
capillary contrast; i.e., the difference in PCE entry pressures (cf. Table 1) for F20-F30 to
F35-F50 sands, and for F35-F50 to F70-F110, sands is 10.12 and 27.27 cm of water
pressure, respectively. In Fig. 4b, the strength of the capillary barrier was sufficient to
allow some of the free product NAPL to cascade around the sides of the F70-F110 lens
and become entrapped as residual NAPL. In contrast, the strength of the capillary barrier
was weaker in Fig. 4a, and PCE penetrated the F35-F50 sand and migrated below the
(F20-F30 to F35-F50) interface.
In Fig. 4c,d, the lenses consist of organic-wet F20-F30 and F70-F110 sand, respec-
tively. The PCE is again at residual saturation directly below the injection point. In
contrast to Fig. 4b, observe in Fig. 4d that the organic-wet F70-F110 sand lens does not act
as a capillary barrier to infiltrating PCE. In Fig. 4c,d capillary forces in the organic-wet
Fig. 3. The spatial distribution of PCE after 360 days when variously wetted F35-F50 sand lenses (1 m high by
2.5 m wide—boundary denoted by the black box) are embedded in the center of water-wet F35-F50 sand. In
(a–e), the organic-wet mass fraction of the sand lens equals 0.0, 0.25, 0.50, 0.75, and 1.00, respectively.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157144
Fig. 3 (continued).
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 145
Fig. 3 (continued).
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157146
lenses actually enhance the infiltration of PCE into these lenses. These same forces create
a very effective capillary barrier, high PCE saturations, and enhanced lateral spreading at
the lower boundary of the lens. Comparison of Fig. 4c,d reveals that higher PCE
saturations occur at the lower lens boundary for the coarser (Fig. 4c, F20-F30) than for
the finer (Fig. 4d, F70-F110) organic-wet sand lens. This observation is somewhat counter
intuitive, since the magnitude of capillary contrast between F30-F50 sand is greater for
F70-F110 than for F20-F30 organic-wet sand. Inspection of Fig. 4d, however, reveals that
the lower permeability and increased magnitude of capillary forces in the F70-F110 lens
(in comparison to Fig. 4c) results in higher PCE saturations throughout the lens and
therefore a lower PCE saturation at the bottom.
In summary, Figs. 3 and 4 demonstrate that the coupled influence of heterogeneity in
soil texture, and wettability controls the formation of subsurface capillary barriers.
Capillary barrier performance can be enhanced or diminished by the presence of subsur-
face wettability variations. For example, Fig. 4a,c indicate that increasing the organic-wet
fraction of a coarser-textured lens results in an increase in strength of the capillary barrier
at the bottom of the lens. In contrast, Fig. 4b,d demonstrate that increasing the organic-wet
fraction of a finer-textured lens results in a decrease in strength of the capillary barrier at its
top surface.
3.2. Dissolution
This section discusses the simulated dissolution behavior of residual NAPL in the
source zone. Before initiating dissolution simulations, two-dimensional simulations of
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 147
PCE infiltration and redistribution were conducted using MVALOR (cf. Figs. 3 and 4). For
natural groundwater flow conditions (no pumping), PCE migration typically occurs over a
much shorter time scale than PCE dissolution. For simplicity and convenience, the
simulated PCE flow and transport processes have decoupled herein. The model domain,
nodal spacing, and boundary conditions for MVALOR simulations were discussed in
Section 3.1. The final organic saturation distribution from MVALOR served as the initial
conditions for the MISER dissolution simulations. A corresponding domain size and nodal
spacing was utilized for MISER simulations. No flow boundary conditions were enforced
at the top and bottom of the MISER model domain, while a 2% hydraulic gradient was
maintained throughout the domain using constant pressure boundary conditions on the left
and right hand sides (flow from left to right). Model simulations were continued until all
the NAPL was completely dissolved (less than 0.003% of the initial NAPL mass
remaining) from the simulation domain.
The MISER code uses nodal specific fluxes, thereby eliminating discontinuities in the
velocity field that are present when using element average velocities. The use of a
continuous velocity field has been shown to yield significant improvements in mass
balance for the transport equation (Yeh, 1981). Temporal error is controlled via limits on
the number of iterations the solver takes per time step. The cumulative mass balance error
(Abriola et al., 1997) for PCE during the dissolution simulations was less than
1.74� 10� 1%.
In addition to the previously presented fluid/matrix properties (density, viscosity,
interfacial tension, and hydraulic property model parameters), relevant PCE transport
properties that were used in the simulations include an equilibrium solubility of
approximately 203 mg l� 1 at 20 jC (Bradford et al., 1999), and an aqueous phase
diffusion coefficient of 6.56� 10� 6 cm2 s� 1 (Hayduk and Laudie, 1974). The horizontal
and vertical dispersivities were selected to be 0.35 and 0.035 m, respectively.
Two-dimensional PCE entrapment and dissolution simulations were conducted for
uniform F35-F50 porous media having organic-wet mass fractions equal to 0.0, 0.25, 0.50,
0.75, and 1.00. To minimize NAPL pooling at the bottom boundary (only the Fo = 0.25
had slight pooling at the bottom boundary), PCE injection was continued for only 0.75
days (half the PCE volume of other simulations). Fig. 5 presents a plot of the average
effluent concentration as a function of water volume exiting along the right boundary.
Since the intrinsic permeability of the F35-F50 sand is independent of wettability, the
volume of water given in the figure can be related to a dissolution time by dividing by the
approximate water flux of 4700 l day� 1. The dissolution time is shortest for the Fo = 0.50
and longest for the Fo = 0.0 system. Intermediate PCE dissolution times occurred in the
other soil wettability systems. Bradford et al. (1999) presented soil column dissolution
data for residual PCE entrapped in fractional wettability F35-F50 sand. For similar NAPL
saturations, these authors observed a trend of decreasing remediation time with increasing
organic-wet mass fraction of the soil. A similar trend is observed in Fig. 5 for PCE
dissolution in soils having Fo equal to 0.0, 0.25, and 0.50. In contrast, PCE dissolution in
soils having Fo equal to 0.75 or 1.00 actually exhibits a slightly longer dissolution time
than the Fo = 0.50 system. An explanation is obtained by considering the spatial
distribution of NAPL in these systems. Similar to Fig. 1, two-dimensional entrapment
simulations reveals that the source zone area (region with So>0.001) increases in the
Fig. 4. A plot of the spatial distribution of PCE after 360 days when various sand lenses (1 m high by 2.5 m
wide—boundary denoted by the black box) were embedded in the center of water-wet F35-F50 sand. In (a) and
(b), the lens consists of water-wet F20-F30 and F70-F110 sand, respectively, whereas in (c) and (d), the lens
consists of organic-wet F20-F30 and F70-F110 sand, respectively.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157148
Fig. 4 (continued).
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 149
Fig. 5. Simulated dissolution behavior for PCE retained in F35-F50 porous media having organic-wet mass
fractions equal to 0.0, 0.25, 0.50, 0.75, and 1.00. The figure presents a semilog plot of the average normalized
effluent concentration (C/Ce) as a function of water volume exiting along the right boundary.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157150
following order: Fo = 0.25>0.50>0.75>1.00>0.0. Dissolution is facilitated in systems with
a higher source zone area due to lower NAPL saturations and a corresponding higher water
permeability. Hence, Fig. 5 suggests that the mass transfer rate (increases with increasing
organic-wet fraction) and distribution are important factors in determining the remediation
time of NAPL contaminants.
To further highlight the influence of NAPL distribution on dissolution time, additional
entrapment and dissolution simulations were conducted for water-wet F35-F50 porous
media having various organic-wet mass fractions (0, 0.25, 0.50, 0.75, 1.00) sand lenses
(1 m high by 2.5 m wide) embedded in the center of the model domain. The PCE
entrapment behavior was presented earlier in Fig. 3. The system is homogeneous for the
Fo = 0 sand lens. Fig. 6 presents a plot of the average effluent concentration as a function
of water volume exiting along the right boundary. The shape of the effluent concen-
tration curve is dependent on the wettability of the sand lens. For relative effluent
concentrations greater than 0.0001, the Fo = 0.50 sand lens system has the slowest
dissolution time, whereas the Fo = 0.75 and 1.00 sand lens systems exhibit slightly faster
dissolution times. In these systems, the spatial distribution of PCE (cf. Fig. 3c,d,e) is
quite similar, and differences in the effluent concentration curves are therefore attributed
to the increased dissolution rate (Bradford et al., 1999) and water relative permeability
(Bradford et al., 1997) with increasing organic-wet sand fractions.
In contrast, Fig. 6 indicates that the homogeneous Fo = 0 system exhibits a more rapid
dissolution behavior than the heterogeneous Fo = 0.5, 0.75, and 1.00 lens systems. Since
the Fo = 0 system has a lower dissolution rate than the Fo = 0.5, 0.75, and 1.00 sands
(Bradford et al., 1999), differences in the effluent concentration curves can be ascribed
primarily to the spatial distribution of PCE. The presence of capillary barriers in the
heterogeneous Fo = 0.5, 0.75, and 1.00 lens systems produce higher PCE saturations and an
associated decrease in water permeability than the homogeneous Fo = 0 system (cf. Fig. 3).
The Fo = 0.25 sand lens system has the highest effluent concentrations initially, achieves
Fig. 6. Simulated dissolution behavior for PCE retained in water-wet F35-F50 sand containing variously wetted
F35-F50 sand lenses (cf. Fig. 3a–e). The figure presents a semilog plot of the average normalized effluent
concentration (C/Ce) as a function of water volume exiting along the right boundary.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 151
relative effluent concentrations lower than 0.00011 the most rapidly, and then exhibits
persistent concentration tailing at low concentration levels. The initial rapid dissolution
behavior for PCE in the Fo = 0.25 sand lens system is due to the lower PCE saturations that
are distributed over a larger region (cf. Fig. 3b), and the enhanced dissolution rate
compared to the uniform water-wet soil. The low concentration tailing behavior occurs
due to the presence of a weak capillary barrier that produces high PCE saturations that
dissolve at a slower rate compared than the Fo = 0.5, 0.75, and 1.00 lens systems.
Comparison of Figs. 5 and 6 reveals that NAPL distribution is very important in
determining dissolution times. For example, the dissolution time for PCE in the
Fo = 0.50 system is shortest in Fig. 5 and longest in Fig. 6.
Two-dimensional PCE entrapment and dissolution simulations were conducted for
water-wet F35-F50 porous media having various water-wet sand (F20-F30, F35-F50, and
F70-F110) lenses (1 m high by 2.5 m wide) embedded in the center of the model domain.
The F35-F50 sand lens system is physically homogeneous. The PCE entrapment behavior
was discussed earlier in Figs. 3 and 4. Fig. 7 presents a plot of the average effluent
concentration as a function of water volume exiting along the right boundary for these
water-wet systems. Observe that the volume of water required to dissolve the PCE
increases in the following order: uniform F35-F50 < F70-F110 lens systembF20-F30 lens
system. Due to the presence of a capillary barrier in the F70-F110 system, Fig. 4b indicates
that PCE is retained exclusively in the F35-F50 sand adjacent to the F70-F110 lens.
Differences in the dissolution behavior of the uniform F35-F50 system and F70-F110 lens
system are therefore primarily due to differences in the spatial distribution of PCE. The
presence of a capillary barrier in the F70-F110 lens system leads to high PCE saturations
(Fig. 4b) that are associated with a reduction in water relative permeability in the source
zone. Furthermore, the lower intrinsic permeability of the F70-F110 lens shields the PCE
on the downstream side of the lens from flowing water. All of these factors account for the
greater volume of water required to dissolve the PCE in the F70-F110 system compared to
Fig. 7. Simulated dissolution behavior for PCE retained in water-wet F35-F50 sand containing water-wet sand
lenses of various soil textures (cf. Figs. 3a and 4a,b). The figure presents a semilog plot of the average normalized
effluent concentration (C/Ce) as a function of water volume exiting along the right boundary.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157152
the uniform F35-F50 system. The F20-F30 lens system requires the largest volume of
water to dissolve PCE in the source zone because of flow bypassing of the PCE retained at
the bottom boundary of the F20-F30 lens. A much greater volume of water flows through
the F20-F30 system due to differences in the intrinsic permeability of the lens sand. The
approximate water flux flowing out of the water-wet F20-F30, F35-F50, and F70-F110
systems is 9700, 4700, and 3800 l day� 1, respectively.
To explore the coupled influence of heterogeneity in soil texture and wettability, two-
dimensional PCE entrapment and dissolution simulations were conducted for water-wet
F35-F50 porous media having various organic-wet sand (F20-F30, F35-F50, and F70-
F110) lenses (1 m high by 2.5 m wide) embedded in the center of the model domain. The
PCE entrapment behavior was discussed earlier in Figs. 3 and 4. Fig. 8 presents a plot of
the average effluent concentration as a function of water volume exiting along the right
boundary for these systems. Observe that the volume of water required to dissolve PCE in
the source zone increases in the following order: F35-F50 < F70-F110bF20-F30 organic-
wet lens systems. Since similar PCE dissolution rates were observed in organic-wet F20-
F30, F35-F50, and F70-F110 sand (Bradford et al., 2000), an explanation of the behavior
shown in Fig. 8 is obtained by considering the aqueous phase flow field and the spatial
PCE saturation distribution. Figs. 3e and 4c,d indicate that each of the organic-wet lens
(F20-F30, F35-F50, and F70-F110) acted as an effective capillary barrier, retaining the
PCE in the lens. The spatial distribution within the lens, however, is not the same. The
F70-F110 lens system has higher PCE saturations distributed throughout the lens than the
other systems. Due to differences in the lens intrinsic permeability, much more water flows
through the F20-F30 (9700 l day� 1) than the F35-F50 (4700 l day� 1) or F70-F110 (3800 l
day� 1) organic-wet lens systems. As for the water-wet lens systems (Fig. 7), the greatest
volume of water is required to dissolve PCE in the F20-F30 organic-wet lens system
because of flow bypassing of the high PCE saturations spread along the bottom boundary
of this lens. The F70-F110 organic-wet lens system exhibits higher effluent concentration
Fig. 8. Simulated dissolution behavior for PCE retained in water-wet F35-F50 sand containing organic-wet sand
lenses of various soil textures (cf. Figs. 3e and 4c,d). The figure presents a semilog plot of the average normalized
effluent concentration (C/Ce) as a function of water volume exiting along the right boundary.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 153
initially and earlier reduction in effluent concentrations due to the more dispersed
distribution of PCE throughout the source zone (cf. Fig. 4). The concentration tailing
behavior of the F70-F110 organic-wet lens system is due to flow bypassing of the high
PCE saturations at the bottom boundary of the lens, which is facilitated by the lower
intrinsic permeability of the F70-F110 sand. Variations in the aqueous phase flow field are
not as pronounced in the F35-F50 organic-wet lens system, and the effluent concentration
curve therefore decreases more uniformly than in the other organic-wet lens systems.
4. Summary and conclusions
Refined conceptual models for NAPL entrapment and dissolution were implemented in
existing two-dimensional NAPL flow and transport simulators. Independently determined
hydraulic properties, residual saturations, and dissolution parameters for soils encompass-
ing a range in grain-size and wettability characteristics were then used in conjunction with
the numerical simulators to explore the influence of heterogeneity in soil texture and
wettability on NAPL entrapment and dissolution.
NAPL infiltration and redistribution simulations demonstrate that the spatial distribu-
tion of residual NAPL (after 360 days) is highly dependent on the subsurface wettability
characteristics. In physically homogeneous systems the final saturation distribution
depends on the values of the residual, entrapped, and immobile (coating NAPL-wet solid
surfaces) NAPL saturations. For a given soil, the NAPL infiltration depth was found to be
maximum at intermediate organic-wet mass fractions (Fo = 0.25). This observation was
attributed to the fact that NAPL entrapment is inhibited by the presence of NAPL-wet
solid surfaces and by associated decreases in the immobile NAPL saturation. Conversely,
NAPL infiltration depth was minimum for strongly wetted soils (water-wet or organic-
wet). This occurs because NAPL entrapment is maximum in water-wet soils, and the
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157154
immobile NAPL saturation is maximum in organic-wet soils. The soil intrinsic and relative
permeability also influence the spatial distribution of NAPL by affecting the historic
maximum NAPL saturation. Decreasing the NAPL conductivity leads to higher NAPL
saturations and, consequently, enhances NAPL retention.
In heterogeneous systems the final distribution of NAPL is also controlled by the
presence of subsurface capillary barriers. Capillary barriers occur as a result of abrupt
changes in capillary characteristics (pore-size distribution or wettability characteristics)
which retain or exclude the NAPL at this interface. NAPL pooling and spreading occurs at
the interface of capillary contrast due to an inhibition of the downward migration of
NAPL. Numerical simulations demonstrate that the performance of soil texture-induced
capillary barriers can be enhanced or diminished by the presence of subsurface wettability
variations.
Numerical simulations were also conducted to explore the dissolution behavior of
residual NAPL in heterogeneous porous media. In physically homogeneous systems, the
dissolution behavior was found to be strongly dependent on the soil wettability, as well as
the spatial NAPL distribution. For similar dissolution rates, NAPL distributions that are
distributed over a larger region are remediated more rapidly because the aqueous phase
relative permeability is greater. In heterogeneous systems, the dissolution behavior was
found to be highly dependent on the spatial NAPL distribution. Spatial distributions of soil
texture and wettability that produced strong capillary barriers and higher NAPL saturations
tended to exhibit longer dissolution times than comparable homogeneous systems.
Numerical simulations suggest a complex interaction between the remediation time and
the NAPL saturation distribution, aqueous phase flow field, and mass transfer character-
istics. The relative importance of each of these factors depends on the degree and type of
heterogeneity.
Acknowledgements
Funding for this research was in part provided by the National Institute of
Environmental Health Sciences (NIEHS) under Grant # ES04911 and by the Department
of Energy (DOE) under Grant # DE-FG07-96ER14702. The research described in this
article has not been subject to NIEHS or DOE review and, therefore, does not ne-
cessarily reflect the views of these organizations, and no official endorsement should be
inferred.
References
Abriola, L.M., 1989. Modeling multiphase migration of organic chemicals in groundwater systems—a review
and assessment. Environ. Health Perspect. 83, 117–143.
Abriola, L.M., Rathfelder, K.M., 1993. Mass balance errors in modeling 2-phase immiscible flows—causes and
remedies. Adv. Water Resour. 16, 223–239.
Abriola, L.M., Rathfelder, K., Maiza, M., Yadav, S., 1992. VALOR code version 1.0: a PC code for simulating
immiscible contaminant transport in subsurface systems. EPRI TR-101018, Project 2879-08. Final Report.
Abriola, L.M., Lang, J., Rathfelder, K., 1997. Documentation for: Michigan Soil-Vapor Extraction Remediation
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 155
(MISER) Model—A Computer Program to Model Soil Vapor Extraction and Bioventing of Organic Chem-
icals in Unsaturated Geological Material. EPA/600/R-97/099. United States Environmental Protection
Agency, Washington, DC.
Anderson, W.G., 1986. Wettability literature survey: Part 1. Rock/oil/brine interactions and the effects of core
handling on wettability. J. Pet. Technol. 38, 1125–1144.
Anderson, M.R., Johnson, R.L., Pankow, J.F., 1992. Dissolution of dense chlorinated solvents into ground water:
1. Dissolution from a well-defined residual source. Ground Water 30, 250–256.
Berglund, S., 1997. Aquifer remediation by pumping: a model for stochastic–advective transport with non-
aqueous phase liquid dissolution. Water Resour. Res. 33, 649–661.
Bradford, S.A., Abriola, L.M., 2001. Dissolution of residual tetrachloroethylene in fractional wettability porous
media: incorporation of interfacial area estimates. Water Resour. Res. 37, 1183–1195.
Bradford, S.A., Leij, F.J., 1995a. Fractional wettability effects on two- and three-fluid capillary pressure–satu-
ration relations. J. Contam. Hydrol. 20, 89–109.
Bradford, S.A., Leij, F.J., 1995b. Wettability effects on scaling two- and three-fluid capillary pressure–saturation
relations. Environ. Sci. Technol. 29, 1446–1455.
Bradford, S.A., Leij, F.J., 1996. Predicting two- and three-fluid capillary pressure–saturation relations in mixed
wettability media. Water Resour. Res. 32, 251–259.
Bradford, S.A., Leij, F.J., 1997. Estimating interfacial areas for multi-fluid soil systems. J. Contam. Hydrol. 27,
83–105.
Bradford, S.A., Abriola, L.M., Leij, F.J., 1997. Wettability effects on two- and three-fluid relative permeabilities.
J. Contam. Hydrol. 28, 171–191.
Bradford, S.A., Abriola, L.M., Rathfelder, K.M., 1998. Flow and entrapment of dense nonaqueous phase liquids
in physically and chemically heterogeneous aquifer formations. Adv. Water Resour. 22, 117–132.
Bradford, S.A., Vendlinski, R.A., Abriola, L.M., 1999. The entrapment and long-term dissolution of tetrachloro-
ethylene in fractional wettability porous media. Water Resour. Res. 35, 2955–2964.
Bradford, S.A., Phelan, T.J., Abriola, L.M., 2000. Dissolution of residual tetrachloroethylene in fractional wett-
ability porous media: correlation development and application. J. Contam. Hydrol. 45, 35–61.
Chatzis, I., Morrow, N.R., Lim, H.T., 1983. Magnitude and detailed structure of residual oil saturation. Soc. Pet.
Eng. J. 23, 311–326.
Dekker, T.J., Abriola, L.M., 2000. The influence of field-scale heterogeneity on the infiltration and entrapment of
dense nonaqueous phase liquids in saturated formations. J. Contam. Hydrol. 42, 187–218.
Dekker, L.W., Ritsema, C.J., 1994. How water moves in a water repellent sandy soil: 1. Potential and actual water
repellency. Water Resour. Res. 30, 2507–2519.
Demond, A.H., Roberts, P.V., 1991. Effect of interfacial forces on two-phase capillary pressure– saturation
relationships. Water Resour. Res. 27, 423–437.
Demond, A.H., Desai, F.N., Hayes, K.F., 1994. Effect of cationic surfactants on organic liquid–water capillary
pressure–saturation relationships. Water Resour. Res. 30, 333–342.
Demond, A.H., Rathfelder, K.M., Abriola, L.M., 1996. Simulation of organic liquid flow in porous media using
estimated and measured transport properties. J. Contam. Hydrol. 22, 223–239.
Essaid, H.I., Herkelrath, W.N., Hess, K.M., 1993. Simulation of fluid distributions observed at a crude oil spill
site incorporating hysteresis, oil entrapment, and spatial variability of hydraulic properties. Water Resour. Res.
29, 1753–1770.
Geller, J.T., Hunt, J.R., 1993. Mass transfer from nonaqueous phase organic liquids in water-saturated porous
media. Water Resour. Res. 29, 833–846.
Guarnaccia, J.F., Imhoff, P.T., Missildine, B.C., Oostrom, M., Celia, M.A., Dane, J.H., Jaffe, P.R., Pinder, G.F.,
1992. Multiphase Chemical Transport in Porous Media, Final Report, EPA/600/S-92/002. U.S. Environ.
Protect. Agency, Washington, DC.
Hayduk, W., Laudie, H., 1974. Prediction of diffusion coefficients for nonelectrolytes in dilute aqueous solutions.
AIChE J. 20, 611–615.
Hunt, J.R., Sitar, N., Udell, K.S., 1988. Nonaqueous phase liquid transport and cleanup: 1. Analysis of mech-
anisms. Water Resour. Res. 24, 1247–1258.
Imhoff, P.T., Jaffe, P.R., Pinder, G.F., 1994. An experimental study of complete dissolution of a nonaqueous
phase liquid in a saturated porous media. Water Resour. Res. 30, 307–320.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157156
Imhoff, P.T., Frizzell, A., Miller, C.T., 1997. Evaluation of thermal effects on the dissolution of a nonaqueous
phase liquid in porous media. Environ. Sci. Technol. 31, 1615–1622.
Klute, A., Dirksen, C., 1986. Hydraulic conductivity and diffusivity: laboratory methods. In: Klute, A. (Ed.),
Method of Soil Analysis, Part 1. Physical and Mineralogical Methods, 2nd ed. Soil Science Society of
America, Madison, WI, pp. 687–734.
Kueper, B.H., Frind, E.O., 1991. Two-phase flow in heterogeneous porous media: 2. Model application. Water
Resour. Res. 27, 1059–1070.
Kueper, B.H., Abbott, W., Farquhar, G., 1989. Experimental observations of multiphase flow in heterogeneous
porous media. J. Contam. Hydrol. 5, 83–95.
Kueper, B.H., Redman, D., Starr, R.C., Reitsma, S., Mah, M., 1993. A field experiment to study the behavior of
tetrachloroethylene below the water table: spatial distribution of residual and pooled DNAPL. Ground Water
31, 756–766.
Land, C.S., 1968. Calculation of imbibition relative permeability for two- and three-phase flow from rock
properties. Soc. Pet. Eng. J. 8, 149–156.
Leverett, M.C., 1941. Capillary behavior in porous solids. Trans. Am. Inst. Min. Metall. Pet. Eng. 142, 152–169.
Mayer, A.S., Miller, C.T., 1996. The influence of mass transfer characteristics and porous media heterogeneity on
nonaqueous phase dissolution. Water Resour. Res. 32, 1551–1567.
Miller, C.T., Poirier-McNeill, M.M., Mayer, A.S., 1990. Dissolution of trapped nonaqueous phase liquids: mass
transfer characteristics. Water Resour. Res. 26, 2783–2796.
Miller, C.T., Christakos, G., Imhoff, P.T., McBride, J.F., Pedit, J.A., Trangenstein, J.A., 1998. Multiphase flow
and transport modeling in heterogeneous porous media: challenges and approaches. Adv. Water Resour. 21,
77–120.
Morrow, N.R., 1970. In: de Wiest, R.J.M. (Ed.), Physics and thermodynamics of capillary action in porous media.
Flow through porous media. Academic Press, New York, pp. 104–128.
National Research Council, 1994. Alternatives for Ground Water Cleanup. National Academy Press, Washington,
DC.
O’Carroll, D.M., Bradford, S.A., Abriola, L.M., 2001. Infiltration and redistribution of PCE in a system con-
taining spatial wettability variations. J. Contam. Hydrol. (submitted for publication).
Parker, J.C., Lenhard, R.J., 1987. A model for hysteretic constitutive relations governing multiphase flow: 1.
Saturation–pressure relations. Water Resour. Res. 23, 2187–2196.
Parker, J.C., Katyal, A.K., Kaluarachchi, J.J., Lenhard, R.J., Johnson, T.J., Jayaraman, K., Jayaraman, K., Unlu,
K., Zhu, J.L., 1991. Modeling multiphase organic chemical transport in soils and ground water. Final Report,
EPA/600/2-91/042. U.S. Environmental Protection Agency, Washington, DC.
Poulsen, M., Kueper, B.H., 1992. A field experiment to study the behavior of tetrachloroethylene in unsaturated
porous media. Environ. Sci. Technol. 26, 889–895.
Powers, S.E., Tamblin, M.E., 1995. Wettability of porous media after exposure to synthetic gasolines. J. Contam.
Hydrol. 19, 105–125.
Powers, S.E., Loureiro, C.O., Abriola, L.M., Weber Jr., W.J., 1991. Theoretical study of the significance of
nonequilibrium dissolution of nonaqueous phase liquids in subsurface systems. Water Resour. Res. 27,
463–477.
Powers, S.E., Abriola, L.M., Weber Jr., W.J., 1992. An experimental investigation of NAPL dissolution in
saturated subsurface systems: steady state mass transfer rates. Water Resour. Res. 28, 2691–2705.
Powers, S.E., Abriola, L.M., Weber Jr., W.J., 1994a. An experimental investigation of NAPL dissolution in
saturated subsurface systems: transient mass transfer rates. Water Resour. Res. 30, 321–332.
Powers, S.E., Abriola, L.M., Dunkin, J.S., Weber Jr., W.J., 1994b. Phenomenological models for transient NAPL-
water mass-transfer processes. J. Contam. Hydrol. 16, 1–33.
Powers, S.E., Nambi, I.M., Curry Jr., G.W., 1998. Non-aqueous phase liquid dissolution in heterogeneous
systems: mechanisms and a local equilibrium modeling approach. Water Resour. Res. 34, 3293–3302.
Rathfelder, K.M., Abriola, L.M., 1998. The influence of capillarity in numerical modeling of organic liquid
redistribution in two-phase systems. Adv. Water Resour. 21, 159–170.
Rathfelder, K.M., Lang, J.R., Abriola, L.M., 2000. A numerical model (MISER) for the simulation of coupled
physical, chemical, and biological processes in soil vapor extraction and bioventing systems. J. Contam.
Hydrol. 43, 239–270.
S.A. Bradford et al. / Journal of Contaminant Hydrology 67 (2003) 133–157 157
Rathfelder, K.M., Abriola, L.M., Taylor, T.P., Pennell, K.D., 2001. Surfactant enhanced recovery of tetrachloro-
ethylene from a porous medium containing low permeability lenses: 2. Numerical simulation. J. Contam.
Hydrol. 48, 351–374.
Rixey, W.G., 1996. The long-term dissolution characteristics of a residually trapped BTX mixture in soil. Hazard.
Waste Hazard. Mater. 13, 197–211.
Saba, T., Illangasekare, T.H., 2000. Effect of groundwater flow dimensionality on mass transfer from entrapped
nonaqueous phase liquid contaminants. Water Resour. Res. 36, 971–979.
Wilson, J.L., Conrad, S.H., Mason, W.R., Peplinski, W., Hagan, E., 1990. Laboratory investigation of residual
liquid organics from spills, leaks and the disposal of hazardous wastes in groundwater. Final Report, EPA/600/
6-90/004. U.S. Environmental Protection Agency, Washington, DC.
Yeh, G.-T., 1981. On the computation of Darcian velocity and mass balance in the finite element modeling of
groundwater flow. Water Resour. Res. 17, 1529–1534.